selecting legal advisors in m&as: organizational … · selecting legal advisors in m&as 2...

Selecting Legal Advisors in M&As 1

SELECTING LEGAL ADVISORS IN M&AS: ORGANIZATIONAL LEARNING AND

THE ROLE OF MULTIPLICITY OF MENTAL MODELS*

Bastian Westbrock

Utrecht University

Katrin Muehlfeld

Trier University

Utrecht University

Utz Weitzel

Utrecht University

Radboud University

FORTHCOMING IN: JOURNAL OF MANAGEMENT

*Supplemental material available online: https://utzweitzel.files.wordpress.com/2017/12/online-

appendix-corr-table-a4.pdf

Acknowledgements: The authors would like to thank associate editor William P. Wan and two

anonymous reviewers for their excellent comments and guidance in improving the article. They

also thank seminar audiences at the AOM 2013 and AIB 2013 conferences for useful comments

on earlier drafts. The authors particularly thank Joanna Barth and Anja Loderer for valuable

research assistance in terms of data collection. Any remaining errors are our own.

Corresponding author: Katrin Muehlfeld, Trier University, Universitätsring 15, 54286 Trier,

Germany. Email: [email protected]

https://utzweitzel.files.wordpress.com/2017/12/online-appendix-corr-table-a4.pdf

https://utzweitzel.files.wordpress.com/2017/12/online-appendix-corr-table-a4.pdf

mailto:[email protected]


ABSTRACT

Management scholars have identified a variety of firm characteristics as antecedents to

organizational learning. In this study, we conceptualize intra-organizational multiplicity of

mental models as a complementary element that facilitates shifting from lower- to higher-level

learning. Specifically, we investigate whether multiplicity of mental models—proxied by four

different measures—, helps acquirers to categorically adapt selection rules for legal advisors in

M&As from domestic towards international settings. In developing our conceptual framework,

we integrate resource-based, social network, and organizational learning perspectives.

Empirically, we draw on 11,511 acquisition attempts announced in 1998–2010 (completion/

abandonment assessed as of January 10th, 2015, at the latest). The results largely support our

theory: First, choosing legal advisors in domestic and international deals calls for different

selection rules. While in domestic deals, network-related characteristics are more important

drivers of lawyers’ performance relative to their country-specific expertise, the comparative

relevance of these attributes is reversed in cross-border deals. Yet, initially, acquirers fail to

recognize this. Also, they do not initially adjust their selection criteria appropriately in response

to accumulating M&A experience of their own. Only after having attempted a substantial

number of cross-border M&As, they reach a turning point at which they rebalance their selection

criteria such that they reflect the predominant relevance of country-expertise in cross-border

settings. Finally, recognition of the need to categorically reassess selection criteria in

international deals is significantly facilitated by an acquirer’s multiplicity of mental models.

Keywords: Organizational learning; experiential learning; social networks; legal advisors; cross-

border vs. domestic M&As; acquisition completion


SELECTING LEGAL ADVISORS IN M&AS: ORGANIZATIONAL LEARNING AND

THE ROLE OF MULTIPLICITY OF MENTAL MODELS

INTRODUCTION

Some organizations appear to be systematically better at learning than others (Argote,

2013). What sets them apart? Argote and Miron-Spektor (2011) conceptualized the

organizational context as composed of various elements whose interplay affects an

organization’s ability to learn. Prior research has empirically identified several contextual

antecedents of organizational learning, such as organizational scope (specialist vs. generalist),

culture, absorptive capacity, aspiration level, intra- and inter-organizational network relations, IT

and knowledge management systems, and, for intra-organizational group learning processes,

lower-level constructs such as team member diversity (Argote, 2013).

Here, we propose intra-organizational multiplicity of mental models—defined as a count

of distinct mental models relating to a specific information domain within the organization—as a

complementary contextual element that may foster specific types of organizational learning. Our

arguments differ from prior research in that we propose that the impact of multiplicity of mental

models is less domain-specific than the influence of related constructs such as absorptive

capacity (Cohen & Levinthal, 1990). We suggest that it allows for ‘meta-contextual’ learning,

that is learning in and across various and even otherwise unrelated domains. In the context of

organizational learning, mental models have been characterized as “deeply held internal images

of how the world works” (Senge, 1990:163). Greater multiplicity of mental models within the

organization, first, makes available a larger range of cognitive structures for describing and

dealing with a specific type of task (e.g., sourcing advisory services). Second, we propose that

multiplicity of mental models facilitates the emergence of a generalized awareness of the


existence of a whole range of cognitive structures—independent of the specific task. By

highlighting, generally, the possibility of ‘doing things differently’, we maintain that multiplicity

may facilitate a specific type of learning: recognition, in response to accumulating experience, of

the need to shift from lower-level learning, which entails adaptation within a given set of rules,

to higher-level, rule-changing learning (cf. Fiol & Lyles, 1985).

In this paper, we thus seek to address whether an organization’s multiplicity of mental

models positively affects recognition of the need to shift to higher-level learning. We address

this issue in the context of a specific research setting: Acquirers’ selection of legal advisors in

mergers and acquisitions (M&As) constitutes the object of learning in this study. Given that the

relevance of different types of advisors varies across the M&A process, we zoom in on a specific

stage during which legal advisors are of particular importance, the public takeover phase

(Karsten, Malmendier, & Sautner, 2014; Krishnan & Masulis, 2013). We study if acquirers apply

selection rules, which they have developed in domestic M&As, to international deals: On the one

hand, institutional differences across countries (e.g., Dikova, Rao Sahib, & van Witteloostuijn,

2010) create strong pressures to re-assess established routines, in particular related to the role of

country-specific legal expertise. On the other hand, the credence good nature of legal services

(Coates, 2001) and the potential ‘dark side’ of embeddedness (e.g., Lee, 2013) of acquirers’

network relations with lawyers counteract these pressures.

Following research on partner selection (e.g., Gulati & Gargiulo, 1999; Hitt, Dacin,

Levitas, Arregle, & Borza, 2000; Lee, 2013), we blend resource-based, social network, and

organizational learning perspectives in deriving our hypotheses. First, we analyze acquirers’

reliance in domestic deals on relational- vis-à-vis expertise-related law firm attributes for advisor

selection and consider their respective contribution to advisor performance. Second, we


investigate whether the same criteria govern advisors’ performance in international M&As and

whether acquirers employ them as a basis for selection. Our basic expectation is that bidders

need to nominate advisors based on a reversed weighting of relational and expertise-related

attributes when they operate internationally, but that they do not recognize this initially. Next, we

study the effect of experiential learning: When accumulating international M&A experience, do

acquirers learn to fundamentally rebalance the selection criteria to better align them with the

drivers of law firm performance in international deals? Following Haleblian and Finkelstein’s

(1999) theory of inappropriate generalization, we expect that they will initially fail to do so, but

that sufficient experience will facilitate the shift from lower- to higher-level learning needed for

a categorical rebalancing of selection criteria. Lastly, we ask whether firms with greater

multiplicity of mental models require less experience in order to reach this turning point.

An analysis of 5,594 domestic and 5,917 cross-border M&A attempts announced in 1998–

2010 supports our theory. First, selecting legal advisors calls for distinct weighting of network-

relative to expertise-related attributes in domestic and international deals. Second, acquirers learn

from experience: There exists a turning point beyond which the typical bidder relies more on

expertise- rather than network-related selection criteria in cross-border deals. Third, multiplicity

of mental models—assessed at various organizational levels and in several domains—appears to

facilitate the required shift from lower- to higher-level learning. In sum, this study provides

important insights regarding organizational learning, social networks, and M&As, and for

managerial practice.

RESEARCH SETTING: THE ROLE OF LEGAL ADVISORS IN COMPLETING M&AS

M&A processes consist of two broad stages: pre-completion, which ends with the official

resolution, and integration (Haspeslagh & Jemison, 1991). Two critical events mark the pre-


completion phase (Boone & Mulherin, 2007): The private takeover period starts with the initial

screening of potential business combinations and ends with the official announcement of a

preliminary merger agreement. The ensuing public takeover phase starts with this announcement

and ends with either completion or abandonment. While a legal advisor might be involved in

drafting the preliminary agreement, the primary duty of the counsel begins after this agreement

is signed and the deal is announced (DePamphilis, 2008). The counsel is then responsible for (i)

obtaining shareholder, regulatory, and third-party consent, and (ii) structuring, drafting, and

completing the definitive agreement of purchase and sale in a way that accommodates the

diverse interests and complies with all relevant regulations of every country where acquirer and

target are listed or incorporated (Gilson & Black, 1996). Lawyers advise clients on

responsibilities in making or responding to offers, act as principal negotiators for legal terms,

provide legal due diligence, draft contracts, ensure that corporation, anti-trust, securities and tax

laws are being followed, and are responsible for M&A related litigation (Krishnan & Masulis,

2013). How can their performance during the public takeover phase be assessed?

While the release of new information may always prompt the necessity to reassess a deal

(Chakrabarti & Mitchell, 2016; Zhou, Xie, & Wang, 2016), prior literature has argued that

during this public takeover stage, completion and, in particular, closing time (i.e., the time it

takes to proceed from the initial public announcement to the definitive agreement) constitute

important outcome variables from the viewpoint of acquirers (Chakrabarti & Mitchell, 2016;

Dikova et al., 2010; Karsten et al., 2014).1 Particularly at this stage, the involved parties need to

resolve issues of legal nature in order to swiftly complete an announced M&A. M&As are

complex legal transactions because of the large number of stakeholder groups that is involved,

including managers, shareholders, banks, employees, and regulatory authorities (Boone &


Mulherin, 2007; DePamphilis, 2008). Against this background, the legal counsel has been

described as a “transaction cost engineer” (Gilson, 1984: 244) who has a major impact on the

likelihood that an agreement is reached and, in particular, on the time it takes to close a deal

(Gilson & Black, 1996; Krishnan & Masulis, 2013; Karsten et al., 2014). Hence, a lawyer’s

contribution to reducing closing time can be viewed as a key indicator of its performance.

THEORY AND HYPOTHESES

Acquirers’ Selection of Legal Advisors and Law Firm Performance in Domestic M&As

How do acquirers select lawyers for the public takeover phase and what drives advisors’

performance? Little direct evidence exists, but research on partner selection (e.g., Hitt et al.,

2000) including alliances (e.g., Chung, Singh, & Lee, 2000; Gulati & Gargiulo, 1999), suggests

at least two categories of selection criteria, which may also affect performance: first, social

network relations; second, access to complementary resources and capabilities.

Social network relations. The literature on partner selection in inter-organizational

networks (e.g., Gulati, 1995; Mitsuhashi & Min, 2016) and on social networks in professional

services (e.g., Baum, Rowley, Shipilov, & Chuang, 2005; Jensen, 2003) has identified network-

related attributes—also referred to as embeddedness—, such as past direct ties between

organizations as strong predictors of selection in future exchange episodes. Social networks

provide unique mechanisms for the governance of transactions and dissemination of information

(Powell & Smith-Doerr, 1994). Ties from repeat interactions between the same actors enable

mutual learning, motivate the sharing of private information, and reduce the potential for

frictions in subsequent episodes (Uzzi & Lancaster, 2004).

Studies on network relations of professional service firms (e.g., Baker, 1990; Uzzi &

Lancaster, 2004) suggest that past direct ties between clients and advisors not only matter for


selection but may also affect outcome variables—as would be expected from a rational

capability-based logic of partner selection. How may network relations between acquirer and

lawyer enable the legal counsel to perform better during the public takeover phase? Past

interactions can help in various ways: First, they enable direct learning about the partner’s

corporate culture and internal procedures. Familiarity with the bidder’s ways of working during

the public takeover stage, with its intra-organizational routines, organizational structure, and

decision-making bodies—both formal and informal—facilitate coordination between lawyer and

client regarding the design of the acquisition contract (Karsten et al., 2014) and, generally, the

structuring of the diverse interests and compliance with all relevant regulations. Second, past ties

facilitate the emergence of trust (Gulati, 1995; Uzzi & Lancaster, 2004); especially if

accompanied by a degree of closeness and exclusivity in the relationship (Coleman, 1990;

Gargiulo & Benassi, 2000) that goes beyond mere familiarity. Trust stimulates relationship-

specific investments, for example, in terms of inter-organizational routines in jointly handling

legal issues during the public takeover stage. Also, it may stimulate sharing of private

information (Gulati, 1995; Uzzi & Lancaster, 2004) about motives of and intentions for a

merger. Such information, in turn, enables the lawyer to better anticipate the acquirer’s concerns

during negotiations, speeding up internal discussions between bidder and advisor and, thereby,

also the process of exchanging mark-ups between the parties (Karsten et al., 2014).

Country-specific expertise. Firms tend to team up with other firms that possess

complementary resources or capabilities (e.g., Chung et al., 2000; Mitsuhashi & Min, 2016; Shah

& Swaminathan, 2008). While much research considers horizontal alliances, the argument likely

extends to acquirers’ relationships with legal advisors. Given the complexity of M&As, which

touch upon many different areas of law (Gilson & Black, 1996), and their relative rarity as


corporate events, most companies find it inefficient to rely strongly on in-house counsel. Instead,

they seek the expertise of law firms specializing in M&As (e.g., Heinz, Nelson, & Laumann,

2001). A broad range of legal domains is relevant in mergers and countries differ significantly in

these respects. Also, given the dispersed and highly regulated nature of the international legal

service industry (e.g., Faulconbridge & Muzio, 2016), law firms’ M&A expertise is often

country-specific: Local experts are often regional or even local players, suggesting (target)

country-specific expertise as a key selection criterion.

Here, the mechanisms that enable a country expert to shorten closing time are particularly

relevant. First, familiarity with the legal and regulatory framework of a focal country implies the

ability to efficiently ensure compliance with all relevant regulations of this country (Gilson &

Black, 1996). Second, routines to apply efficient information verification techniques, which may

vary by country, allow for expediting legal due diligence, among others, because familiarity with

the specific context facilitates interpretation and an assessment whether and what additional

information needs to be collected from local information providers. Third, as suggested by

studies on cultural influences on negotiations, dealings with the seller and its legal counsel are

likely to be influenced by country context (e.g., Adair & Brett, 2005). Having routines in place to

carry out these negotiations should reduce completion time. Fourth, established relationships

with local authorities are an important form of social capital. They facilitate interpretation and

appraisal of regulatory requirements and preparation of necessary documents, thereby reducing

the time that the deal is being reviewed by the authorities. Overall, a law firm’s country-specific

legal expertise likely allows it to engage more effectively in ‘engineering’ a deal. Studies in

financial economics have indeed shown how bidders may benefit from hiring lawyers with


superior general and, to some extent, country-specific expertise during the final phases of

M&As, especially in terms of quicker closing (Karsten et al., 2014; Krishnan & Masulis, 2013).

Which of these lawyer characteristics—network relations or country-specific expertise—

is more important for an acquirer’s selection of a legal counsel; and which is a better predictor of

the counsel’s performance? For domestic M&As, we suggest that country-specific expertise is

less important in both respects: Acquirers generally face difficulties in assessing the quality of an

unknown lawyer, both ex-ante and ex-post (Coates, 2001). Heterogeneity among domestic law

firms in terms of their country-specific expertise in the domestic context is, by definition,

limited. In addition, acquirers are themselves familiar with the basic tenets of the legal

framework governing business operations in their home country. Therefore, we expect acquirers

to rely on their trusted counsel, as they stand to gain little from appointing an unfamiliar one; and

this choice will be in line with the drivers of lawyers’ performance. Overall, we expect:

Hypothesis 1a (H1a): In domestic M&As, network- and expertise-related attributes of a

law firm are positively associated with the likelihood of being selected as counsel, but

acquirers place relatively more weight on network-related criteria.

Hypothesis 1b (H1b): In domestic M&As, network- and expertise-related attributes of a

law firm are positively associated with its performance in terms of shorter closing time,

but network-related criteria have a relatively stronger impact on performance.

Acquirers’ Selection of Legal Advisors and Law Firm Performance in International M&As

First, selection: Do bidders transfer domestic advisor selection routines to the

international domain? Prior research offers contrary perspectives. Some scholars have argued

that acquirers in international deals are acutely aware of their unfamiliarity with the deal context,


including the institutional environment (e.g., Barkema & Schijven, 2008; Muehlfeld, Rao Sahib,

& van Witteloostuijn, 2012), suggesting an increased sensitivity to the need to adapt.

Yet, collecting information about competencies and reliability of another firm is costly

and time consuming (e.g., Stinchcombe, 1990). For legal services, this challenge is aggravated

by the fact that “legal advice often involves questions of judgment under conditions of

uncertainty that will persist even after a trial or negotiation or other legal event is completed”

(Coates, 2001: 1301). Hence, as acquirers face difficulties in assessing the quality of an unknown

(host country) lawyer vis-à-vis the proven qualities of a known advisor, they may transfer the

network-heavy weighting of selection criteria from the domestic to the international domain.

Research on social networks supports this view, pointing to “embeddedness as a solution to

relational hazards in market exchanges” (Lee, 2013: 1238). In the very situations that require

specific expertise to handle a complex task with uncertain outcomes, actors often resort to what

they perceive as the least risky option: their well-known, trusted partners (Lee, 2013; Mizruchi &

Stearns, 2001). This is not to say that acquirers necessarily select their counsel based on

irrational motives. Rather, the uncertainty they face in the nomination decision, combined with

the complexity and uncertainty of the cross-border deal itself, may make them highly risk averse,

‘nudging’ them away from the local expert, and towards the familiar counsel: “when trustworthy

partners are already available, searching for strangers is hard to be rationalized” (Chung et al.,

2000: 5). Indeed, Howard, Withers, Carnes, and Hillman (2016) found that greater firm-level

uncertainty led firms to reinforce alliance ties rather than engage in a broadening of ties. Thus:

Hypothesis 2a (H2a): In cross-border M&As, network- and expertise-related attributes of

a law firm are positively associated with the likelihood of being selected as counsel, but the

typical acquirer places relatively more weight on network-related criteria.


Second, performance: How could target country-specific expertise reduce closing time?

Essentially, we expect the same mechanisms to apply as proposed above: familiarity with the

legal framework, routines for information verification and legal due diligence, familiarity with

culture-specific negotiation styles and customs, established relationships with authorities.

However, compared to domestic M&As, acquirers are up against greater challenges in

attempting to complete international deals due to, among others, national differences in terms of

formal and informal institutions (Dikova et al., 2010; Weitzel & Berns, 2006), such as nation-

specific legal and tax systems and accounting practices (Very & Schweiger, 2001). This

complication affects, in particular, legal issues during the closing phase. Not only does the legal

service industry feature a high degree of national embeddedness (Faulconbridge & Muzio, 2016),

legal advisors additionally face the task of ensuring legal compliance in a highly fragmented

environment with increasingly complex national takeover regulations (White & Case, 2009).

Thus, the provision of legal services in M&As has only limited potential for developing any

generalized expertise (cf. Karsten et al., 2014). In addition, the limitations to transferability of

local expertise are exacerbated by the fact that many countries have installed widespread

impediments to the entry of foreign lawyers. Hence, a lawyer’s local expertise in a core country

of its operations constitutes a key asset in advising deals that involve this country. In sum, we

expect that a law firm’s target country-specific expertise allows it to engage more effectively in

“engineering” a deal including quicker legal due diligence, reduced time for drafting and revising

the contract and mark-ups, more efficient negotiations with the counter-party, and shorter time

that the deal is being reviewed by authorities. Given the eminent role of legal contingencies

during the public stage, we further maintain that in cross-border deals, target country expertise is

even more significant for advisor performance than a trustful relationship and reduced relational


frictions between bidder and advisor. Overall, we expect that the relative importance of network-

and expertise-related attributes is reversed, compared to domestic deals:

Hypothesis 2b (H2b): In cross-border M&As, network- and expertise-related attributes of

a law firm are positively associated with its performance in terms of shorter closing time,

but expertise-related criteria have a relatively stronger impact on performance.

In sum, while H1a-H2a fit into an (at least boundedly) rational capability-based logic of

partner selection, H2b breaks this logic: In cross-border deals, acquirers continue to rely on their

long-term partners, with ties initially established in domestic settings, but these embedded

lawyers are unlikely to be experts in the specific host country environment and vice versa (cf.

Mitsuhashi & Min, 2016). Social network research provides a possible rationale for such a

discrepancy between drivers of selection and of law firms’ performance in cross-border M&As:

Acquirers may favor familiar choices over the most capable ones, because they fall prey to a

behavioral bias arising from a dark side of embeddedness (e.g., Lee, 2013). Can experience

eliminate this discrepancy by stimulating learning? And, if so, could multiplicity of mental

models within the acquirer firm facilitate this learning process?

Types of Organizational Learning and Selection of Legal Advisors in International M&As

Organizational learning refers to an experience-driven change in an organization’s

knowledge, which includes both declarative and procedural knowledge, that is, facts as well as

skills and routines, thus embracing both cognitive and behavioral elements (e.g., Argote, 2013;

Fiol & Lyles, 1985). Routines represent collective, repetitive, and fairly stable patterns (Nelson

& Winter, 1982) that may relate to operational, administrative, and strategic activities, such as,

for example, procedures to select advisory services in M&As. By performing a task, an

organization accumulates experience, giving rise to learning.


Yet, there is more than one kind of learning. Different kinds of learning appear to co-

exist in organizations (e.g., Miller, 1996) and most scholars agree on the existence of at least two

types of qualitatively distinct learning (cf. Tosey, Visser, & Saunders, 2012). Tosey et al. (2012)

identify as the most common labels to capture this dichotomy: lower- and higher-level learning

(Fiol & Lyles, 1985), single- and double-loop learning (Argyris & Schön, 1978), exploitation

and exploration (Levinthal & March, 1993; March, 1991), incremental and radical (Miner &

Mezias, 1996), and adaptive and generative (Senge, 1990). Further, they suggest that a

“consensus seems to have been established that they refer to comparable learning processes and

outcomes” (Tosey et al., 2012: 292). Here, we use the terms lower- and higher-level learning.

Lower-level learning occurs within a given set of rules, norms, and frameworks, is based on

existing routines, maintains the core features of the relevant ‘theory-in-use’, and restricts “itself

to detecting and correcting errors within that given system of rules” (Fiol & Lyles, 1985: 808;

referring to Argyris & Schön, 1978). Higher-level learning implies changing underlying rules,

norms, and frameworks, and aims at developing “new cognitive frameworks within which to

make decisions” (Fiol & Lyles, 1985: 808). It requires “unlearning” of conceived knowledge and

reorientation of cognitive structures (Nystrom & Starbuck, 1984).

Acquirers’ Adaptation of Selection Criteria from Domestic to International Settings

Applied to our research setting, we characterize lower-level learning as attempts to

improve the selection of legal counsels in cross-border M&As without questioning the basic

selection rules, that is, for example, by adapting the weights attached to the two different kinds

of selection criteria transferred from the domestic context, but without categorically changing the

predominant importance of network-related attributes. Based on H1-H2, we argue, though, that

acquirers ultimately need to recognize that the relative significance of network- vis-à-vis


expertise-related attributes is reversed in cross-border M&As. This requires a switch from lower-

to higher-level learning.2 Are acquirers able to use their M&A experience to achieve this shift?

The value of experience for performance-enhancing learning crucially depends on the

extent to which it applies to a future activity (March, 1991), i.e., its transferability is bound by its

context-specificity. Prior studies have emphasized the limits of such transferability and, thus, of

the generalizability of prior experience (e.g., Haleblian & Finkelstein, 1999). Discriminating

between appropriate and inappropriate generalization appears to be particularly difficult in

complex corporate domains due, among others, to causal ambiguity (Argote, 2013). Firms may

need to accumulate a lot of experience before performance improves and, initially, more

experience may even reduce performance due to inappropriate generalization (e.g., Duijsters,

Heimeriks, Lokshin, Meijer, & Sabidussi, 2012; Haleblian & Finkelstein, 1999). Individuals as

well as organizations tend to suffer from resistance to adaptation, due, for example, to cognitive

biases (e.g., Samuelson & Zeckhauser, 1988) and greater uncertainty of exploring new

alternatives vis-à-vis refining existing ones (March, 1991). When the required learning entails a

shift from lower- to higher level, it is likely to be particularly challenging, as it requires ‘distant

search’, while the initial response to unsatisfactory outcomes tends to be ‘local search’ (Cyert &

March, 1963), which may result in further suboptimal outcomes (Barkema & Schijven, 2008).

When acquirers gain experience with selecting advisors, the repeated experience of prior

knowledge (about selection criteria) conflicting with to-be-learned information (antecedents to

advisor performance in cross-border deals), may give rise to the gradual emergence of novel,

alternative cognitive structures for how to select the most suitable legal advisor for a particular

type of deal—domestic or cross-border. Yet, it may take many M&As to achieve this. Thus, we

expect that acquirers’ initial learning will result, if anything, in greater weight of past ties as a


selection criterion, relative to (target) country expertise; and that they need to accumulate a lot of

experience before adequately appreciating a law firm’s (target) country expertise. The turning

point marks, as we argue, the sketched shift in selection routines.

Hypothesis 3 (H3): With growing cross-border M&A experience, acquirers first attach

more relative weight to network-related criteria and, after a turning point, increasingly

attach less weight to network-related relative to expertise-related criteria.

The Impact of Intra-Organizational Multiplicity of Mental Models

Prior research has identified three major types of moderators of the relationship between

experience and learning outcomes (Argote, 2013), that is, types of experience (e.g., direct vs.

indirect), attributes of the learning process (e.g., its mindfulness), and elements of the

organizational context (e.g., organizational culture). Adding to the third domain, we propose that

structural attributes of organizational cognition—in particular, multiplicity of mental models—

may foster a specific aspect of organizational learning, i.e., the ability to recognize, based on the

accumulation of experience, the need to shift from lower- to higher-level learning.

Mental models have initially been conceptualized at the individual level. Rouse and

Morris (1986: 360) characterized a mental model as a “mechanism whereby humans generate

descriptions of system purpose and form, explanations of system functioning and observed

system states, and predictions of future system states.” As such, mental models represent

“organized knowledge frameworks that allow individuals to describe, explain, and predict

behavior” (Lim & Klein, 2006: 404). Senge (1990: 163) characterized them as “images that limit

us to familiar ways of thinking and acting.” Similar to other related concepts (e.g., cognitive

map; for a review, see Walsh, 1995), a mental model thus constitutes a form of knowledge

structure or schema—a fundamental top-down information processing construct characterized by


Walsh (1995: 281) as a “mental template[s] that individuals impose on an information

environment to give it form and meaning”. Scholars have subsequently transferred the

knowledge structure construct to more aggregate levels and have, using a variety of labels,

discussed the existence of supra-individual knowledge structures at the team/group, organization,

and even industry level of analysis (cf. Walsh, 1995, for a review). For example, in the context of

product diversification, Prahalad and Bettis (1986: 491, 485) characterized dominant logic as a

“mind set or worldview or conceptualization of the business and the administrative tools […]

stored as a shared cognitive map”, which “consists of the mental maps developed through

experience in the core business”. Based on these studies, we define mental models at the level of

groups within the organization as collective organized knowledge frameworks that allow for

describing, explaining, and predicting behavior related to a specific information domain. For

example, we expect business units operating in different product markets to each be guided by

such a collective knowledge framework that allows the unit to describe, explain, and predict

behavior in its sector, related, for example, to industry conduct.3

It follows from the characterization of mental models as organized knowledge

frameworks that they have both content and structure (Walsh, 1995). Attributes relating to both

of these dimensions are thought to impact the consequences of these knowledge structures in

guiding future behavior. In terms of content, for example, knowledge structures that represent an

information environment build on experience in this domain, and the context in which this

experience was gained, constitutes an important boundary condition (e.g., Kaplan, 2008; Walsh,

1995). The emergence of knowledge structures appears to entail reflection upon and

interpretation of this experience, including the development of perceptional filters that affect

subsequent information acquisition and processing (e.g., Simons & Chabris, 1999). Here, we


focus solely on structure, though; and, unlike most prior studies concerned with structural

attributes, we consider the structure of a whole set of mental models rather than the structure of

an individual schema. Prior research on schema structure has analyzed differentiation, i.e., the

number of a knowledge structure’s internal dimensions, and integration, i.e., the extent of

interconnectedness of these dimensions (see Walsh, 1995, for a review). Using the term

‘multiplicity’ we consider differentiation at the level of a set of knowledge structures (e.g., the

number of group-level mental models associated with the product markets served by an

organization), a structural attribute that is also referred to as ‘richness’ in organizational ecology

(e.g., Boone, Wezel, & van Witteloostuijn, 2013). Richness captures a pure count of the number

of distinct types in a population or system. Accordingly, multiplicity of mental models refers to a

count of the number of distinct mental models relating to a given information domain within an

organization. Thereby, our definition of multiplicity is both related to and yet distinct from

diversity as commonly conceptualized in management research (see Harrison & Klein, 2007, for

a review). Harrison and Klein (2006) outlined three major ways in which the diversity construct

is being used, that is, diversity as separation, disparity, and variety. The latter is closest to our

definition of multiplicity. Yet, diversity as variety typically takes into account both richness and

abundance or relative distribution of these types (i.e. the proportion of unit members who belong

to a particular type). Thus, multiplicity is a simpler concept, without specific assumptions on

more complex distributional characteristics.

In terms of influencing an acquirers’ ability to shift from lower- to higher level learning,

we argue that multiplicity of mental models leverages the organization-level “ability to reflect

on, understand, and control” (Schraw & Dennison, 1994: 460) the entity’s learning, in particular,

by facilitating the formulation of strategies for considering alternative cognitive mechanisms and


for deliberately selecting from a set of available mechanisms those that are to be applied in a

specific situation (Flavell, 1987). Thus, we propose that multiplicity enables firms to better

utilize their (domain-specific) experience as input for their learning by facilitating more accurate

reflection on the generalizability of this experience: First, multiplicity of mental models relating

to a certain domain may have task- or domain-specific effects—akin to potentially positive

effects from diversity of experience (e.g., Haunschild & Sullivan, 2002; Nadolska & Barkema,

2014)—by making available a range of cognitive structures for describing and dealing with a

specific type of task (e.g., selecting lawyers). By allowing for consideration of a larger range of

perspectives, qualitatively better decisions may become possible (e.g., Bell, Villado, Lukasik,

Belau, & Briggs, 2011; Dahlin, Weingart, & Hinds, 2005), with positive effects of heterogeneity

potentially resulting from superior information processing related to range, depth, and integration

of information use.

Second, resting on our conceptualization of multiplicity of mental models as a content-

independent, structural attribute, we propose that it may additionally stimulate a generalized

awareness effect independent of the specific task or domain. Each mental model rests on a set of

assumptions, rules, and norms. Greater multiplicity of mental models illustrates the possibility to

view the world through a larger number of sets of rules. By raising awareness of the existence of

many distinct worldviews, such multiplicity may facilitate a specific aspect of organizational

learning, that is, recognition of the need to switch from lower- to higher-level learning that

implies changing the underlying rules. Thereby, multiplicity of mental models could allow for

more accurate reflection on the categorization, transferability, and generalizability of

experience—processes that are fraught with difficulties, as prior studies have emphasized (e.g.,

Haleblian & Finkelstein, 1999). Based on the assumption that the co-existence of a multitude of


worldviews in a domain is widely known throughout the organization, acquirers with greater

multiplicity of mental models should thus require less experience to move from lower- to higher-

level learning. Hence, our arguments as to this generalized effect imply that we expect the

beneficial effects of multiplicity to apply beyond a specific task, ranging from closely related

domains to fairly distant and even unrelated ones (for a similar argument, see Powell & Rhee,

2016). What matters are not actual similarities between two tasks, but demonstration of the—at

least hypothetical—possibility to ‘do things differently’. Thus, while experience represents a

major building block for knowledge structures such as mental models (Walsh, 1995), knowledge

about the co-existence of worldviews related to a specific domain (e.g., product markets

associated with business units) may permeate the organization in a way that knowledge about

specific experiences in this domain could not. Thereby, the proposed generalized effect differs

from concepts such as knowledge transfer between units, which often suffers from stickiness that

impedes the reuse of practices in other subunits (Szulanski, 1996). Overall, we expect:

Hypothesis 4 (H4): Multiplicity of mental models is positively associated with an

acquirer’s experiential learning: Acquirers with greater multiplicity of mental models

require less cumulative cross-border M&A experience to reach the turning point beyond

which they attach relatively less weight to network- vis-à-vis expertise-related criteria.

METHODOLOGY

Data and Sample

We retrieved data from the Thomson Reuters’ Financials SDC M&A database

(henceforth: Thomson). It records all publicly announced M&As worldwide, including details on

all legal counsels involved in the transactions. Our initial Sample A consists of 153,622 domestic

and international M&A attempts announced in 1994-2010. We used the available information on


acquirers and counsels to construct our main variables, whereby our definition of acquirer and

counsel coincides with the variables ‘acquirer ultimate parent’ and ‘legal advisor’ in Thomson.

We then reduced Sample A to a smaller set (5,594 domestic; 5,917 cross-border M&As) for our

hypotheses tests. This Sample B includes acquirers and deals (a) that are neither self-tenders nor

minority deals, (b) that have at least one law firm on board, (c) with acquirer and target primarily

active in the non-financial non-legal service industries, (d) where the acquirer attempted at least

one M&A in the four years prior to the focal deal, and (e) where all relevant variables are filled.4

Dependent Variables

For testing H1b and H2b, we generated for each M&A in Sample B the integer variable

duration of the closing phase. It measures the difference between the dates of initial public

announcement and of unconditional completion for every merger officially consummated before

January 11th, 2015. Because durations range from 1 to 3213 days, with a median of 53 days, we

included their natural logarithm, thereby alleviating concerns that the results might be driven by

a small number of deals with a very long closing phase. However, the correct unit of analysis for

testing H1b and H2b is not a merger, but an advisor-acquirer dyad. In order to avoid the

statistical interdependency between the considerable number of M&As in Sample B (34%),

where the acquirer employed a team of two or more advisors, we focused on the ‘lead advisor’,

i.e., the largest lawyer in terms of law firm worldwide experience (variable introduced below).5

For testing H1a, H2a, H3, and H4, we constructed for each M&A in Sample B a ‘risk set’

of potential legal counsels. Within this set, we examined the probability that a given law firm

was selected by the focal acquirer, measured as the likelihood of selection, which is 1 for every

law firm that was appointed (0 otherwise). The risk sets include, next to the actually appointed

counsels, also lawyers that could have potentially been selected. As there are many firms that are


obviously implausible candidates, e.g., a German lawyer in a U.S.-Japan deal, we constructed

sets of ‘shortlisted’ candidates, that had counseled in at least one other deal in the year of the

focal M&A’s announcement, and were among the top 5 percent of firms in terms of at least one

of the four network- and expertise-based law firm attributes described below. A typical risk set in

our analysis then consists of, on average, 104 law firms per deal. Multiplied by the 11,511 deals

in Sample B, the total risk set (Sample C) comprises 1,160,232 acquirer-law firm dyads.

Independent Variables

Advisor-client relationship. We measured the intensity of the law firm-acquirer

relationship by prior interactions between them. We counted for each law firm-acquirer dyad in

Samples B and C the number of (prior) advisor-client ties as recorded in Thomson and interpret

this, in line with Gulati (1995), as a measure of inter-organizational familiarity. Further, we

divided this count measure by the total number of law firms an acquirer had previously worked

with to obtain the percentage variable advisor-client tie strength. Following network literature

(e.g., Jensen, 2003; Mizruchi & Stearns, 2001), this measure serves as an indicator of inter-

organizational trust, as it captures the degree of ‘exclusivity’ of a relationship. The data—as for

all law firm-related variables—was retrieved from Sample A. We counted all matches of the

same name pair in the transactions prior to a focal M&A that precede the deal by no more than

four years. Prior studies on organizational networks have often used such a moving time window

(e.g., Gulati & Gargiulo, 1999; Lee, 2013). For the legal counsel market, it is further justified by

high turnover of personnel, especially in the elite of the profession (Heinz et al., 2001).

Law firm country-specific expertise. We operationalized country-specific expertise by

law firm acquirer country expertise and law firm target country expertise (henceforth acquirer

country expertise and target country expertise). Each variable counts a firm’s prior M&A


advisory appointments in the home country of the focal acquirer and the focal target,

respectively. Using transaction history as expertise measure has several advantages over

alternatives; in particular, as it captures, next to knowledge embedded in the firm, also its

business experience. Uzzi and Lancaster (2004), for example, measure legal expertise of a law

firm directly by weighting the law school degrees of lawyers in each of a firm’s (foreign) offices.

Yet, as expertise is rewarded with more offers, the number of past appointments is a viable proxy

for these more direct dimensions of legal expertise (Lazega, 2001). At the same time, it portrays

a more accurate picture of the service quality, as the transaction history also captures experience

and routines developed through actual practice of the law. We calculated acquirer and target

country expertise as the percentage shares of the total number of a law firm’s past appointments

in the four years prior to a deal. A value of 100% (0%) reflects high (low) expertise in the legal

framework of a country. In this way, the variables capture those elements of service quality that

are difficult to transfer across national contexts. Only very few large law firms are truly

international: Linklaters (U.K.) and Clifford Chance (U.S.), for example, advised foreign clients

in 62%, and 75% of their deals during our sampling period, respectively. The vast majority of

firms, even among the top tier, focus almost entirely on a domestic clientele. By controlling for

overall transaction history, this percentage measure thus reflects the high degree of geographical

specialization in the industry.

Acquirer cross-border M&A experience. Acquirer cross-border M&A experience counts,

for every acquirer in Sample B, all prior cross-border M&As attempted by the firm in the four

years preceding the focal deal. For testing H3, the variable is included in the legal advisor

selection models, where it is interacted with the law firm variables described above. As learning


might be non-linear (e.g., Haleblian & Finkelstein, 1999), we also included the square of cross-

border M&A experience in the selection models.

Acquirers’ multiplicity of mental models. Measuring multiplicity of mental models is

challenging, especially in large-scale research settings, which allow only for indirect assessment

(for direct methods with smaller samples, see, e.g., Kaplan, 2008; Markoczy, 2001). Thus, we

used several measures, which vary along two dimensions. First, in terms of level of analysis, we

constructed three measures at the organizational level and one at the level of the top management

team (TMT). Second, regarding distance from the object of learning, we included several

corporate domains, ranging from more to less distant from selecting legal advisors in M&As.

At the organizational level, we assessed multiplicity of mental models (mmm), first, in

the domain of antitrust systems (e.g., Glendening, Khurana, & Wang, 2016). We classified all

countries in Sample A according to the Wiki project “Antitrust around the World” (Hylton &

Deng, 2007). This Wiki identifies core dimensions of antitrust systems and cardinally ranks

countries’ current and historic systems in terms of the scrutiny they exercise in each of them. We

used the overall antitrust system score. As the overall country scores are clustered between 17

and 25 points, we grouped countries by score percentiles to obtain a more discriminatory

classification. Our theory emphasizes the impact of the simple prevalence of a range of distinct

mental models rather than trying to weight them according to their frequency or relatedness—as

is often done in studies focusing on diversity of experience (e.g., Haunschild & Sullivan, 2002;

Powell & Rhee, 2016). In contrast to such studies, we therefore constructed a simple count

measure (mmm antitrust systems) of the number of different antitrust score percentiles an

acquirer had been exposed to during the four years prior to a focal deal (similarly, see, e.g.,

Dahlin et al., 2005).


Second, as M&As are frequently subject to political manoeuvring (Serdar Dinc & Erel,

2013), we measured multiplicity of mental models in the domain of political systems (mmm

political systems). We classified the countries in Sample A according to the ‘political stability’

indicator of the Worldbank Worldwide Governance Indicators database. The index is a cardinal

number ranging from -2.5 to 2.5 and is available as of 1996. Similar to the mmm antitrust

systems measure, we counted for each acquirer the number of different political systems’ score

percentiles that the firm had been exposed to in the four years prior to the focal deal, which

means that the variable is available for deals announced in 2000-2010. Both of these measures

have the advantage that they depict the focal structural attribute of an organization’s mental

models (i.e., their multiplicity) in a domain that is closely related to the object of learning, i.e.,

selecting legal counsels in cross-border M&As. Yet, for this very reason, it is difficult to

disentangle their effect from the impact of experience (in terms of both volume and diversity) in

these domains. In fact, experience constitutes a major building block of the emergence and

development of any knowledge structure and thus, also a key driver of multiplicity of mental

models. Therefore, in order to test H4, we also sought to assess mental models in less related

corporate domains that bear no immediate relation with the selection of legal counsels in M&As.

Thus, third, we considered the domain of product markets. Prior research has argued that

product markets or industry sectors are associated with distinct, industry-specific shared concepts

of business conduct (Spender, 1987), which may be associated with distinct mental models (e.g.,

Ginsberg, 1989, on the impact of diversification on the content of managers’ mental models).

Various ways exist to assess the degree to which firms have operations in more than product

market, such as the Herfindahl index (e.g., Wan, Hoskisson, Short, & Yiu, 2011). Again, in line

with our theory, we instead used a simple count of the total number of distinct 4-digit SIC codes


in which an acquirer was active at the time of announcing the focal deal. Thomson records for

both acquirer and target in every merger all their industry activities as measured by SIC codes.

Since our unit of analysis is the acquirer’s ultimate parent, we aggregated the subsidiary SIC

codes and counted for every ultimate parent in Sample B the number of unique SIC codes of all

affiliations with the parent at the time of announcing the focal M&A and in the prior four years.

The result is the time-varying variable acquirer mmm product markets.

Fourth, finally, we included a measure of multiplicity related to a firm’s TMT. We

measured the number of nationalities represented among the top decision-makers, divided by the

size of the TMT. Upper Echelon (UE) studies (e.g., Bromiley & Rau, 2016; Hambrick,

Davidson, Snell, & Snow, 1998; Nielsen & Nielsen, 2013) and team studies in general (e.g.,

Dahlin et al., 2005) suggest that nationality/national origin may be particularly strongly related to

the worldviews of decision-makers and the mental models they hold (see, related, Marano,

Arregle, Hitt, Spadafora, & van Essen, 2016; although this view is not uncontested, e.g.,

Markoczy, 1997; for details, see online appendix 1). Note that our operationalization differs from

the way in which UE and team studies usually assess team composition, often based on the Blau

index (e.g., Carpenter & Frederickson, 2001; in relation to TMT diversity and acquisition

learning, e.g., Nadolska & Barkema, 2014), as we are interested purely in how many distinct

mental models are present in a TMT, rather than in any group dynamics. As TMT data are not

recorded in Thomson, we collected this information from archival sources. This, though,

required us to focus, in line with extant UE studies (e.g., Carpenter & Frederickson, 2001), on a

smaller sample of deals and associated acquirers. In order to keep the data collection tractable,

while maximizing the chances of finding the required TMT information, including as well non-

U.S. firms, we zoomed in on the year 2010 and on the largest 50 acquirers that appeared at least


once during that year. Following prior studies (e.g., Carpenter & Frederickson, 2001), we

operationalized the TMT as the top two tiers of management: CEO, members of the Executive

Board (e.g., chief financial officer) and the next-highest management tier including other senior

managers listed as executive officers. We obtained information from firms’ annual reports, SEC

10-K Filings, and, if otherwise not available, from company websites and public sources such as

directorstats.co.uk. Thereby, we were able to acquire complete information for 30 firms involved

as acquirers in 65 cross-border M&As from among our sample in 2010.

Control Variables

In terms of lawyer attributes, based on the close link between service quality and market

rewards (Lazega, 2001), law firm worldwide experience captures general quality differences

between firms. In a service industry such as the legal sector, it also proxies for firm size. Number

of advisors controls for the influence of the size of the advisory team (Hunter & Jagtiani, 2003).

Considering acquirer attributes, as M&A experience has been shown to also have a direct,

potentially non-linear, influence on pre-completion outcomes (e.g., Dikova et al., 2010), we

included an acquirer’s cross-border M&A experience and the squared experience as control

variables in the outcome analyses. For the same reason, we also included the mental model

variables and acquirer country experience and target country experience in these analyses. We

constructed the latter variables in the same way as the corresponding law firm variables.

Acquirer network size counts the number of different law firms a bidder has previously worked

with. Together with advisor-client tie strength, this variable defines an acquirer’s ‘market

interface’ (Baker, 1990) with the legal industry. Public acquirer indicates whether an acquirer is

publicly listed (1/0 otherwise) and acquirer size measures the total asset value of an acquirer and

its subsidiaries (in U.S.$). In terms of transaction attributes, based on prior research (e.g., Wong


& O’Sullivan, 2001), we included public target (1/0 otherwise); deal value (in U.S.$); number of

bidders; the percentage variable stock payment measuring the acquirer’s intended share of stock

payments; and tender offer to identify deals where the acquirer approaches the target’s

shareholders directly (1/0 otherwise). We controlled for timing of a deal by including year

dummies. Based on La Porta, Lopez-de-Silanes, and Shleifer (2008), we controlled for regulatory

hurdles by adding a common law origin dummy if the target country has a common law origin,

i.e. a more merger friendly control regime (Rossi & Volpin, 2004) (1/0 otherwise). Finally, same

legal origin indicates acquirer-target country pairs with the same legal origin (1/0 otherwise).

Estimation Techniques

The major issue with our simultaneous estimation of selection criteria and outcomes is

that the law firms in Sample B are not randomly allocated to deals. Instead, they are selected by

the decision of an acquirer who chooses, as hypothesized, only the most suitable candidates from

Sample C. This may confound the true effect of a law firm’s network- and expertise-based

attributes in the outcome analysis (Cameron & Trivedi, 2005). Even though this problem is

partially alleviated by including various control variables, which together account for the likely

fact that a more difficult case calls for a more competent lawyer, several dimensions of

transaction complexity that we do not observe are likely to remain. To account for these

unobserved dimensions, we estimated Heckman two-stage models (Heckman, 1979) with the

likelihood of selection as the dependent variable in the first stage and duration in the second

stage of our regressions. These models require some regressors in the selection equation that are

not included in the outcome equation, but none of the aforementioned variables qualifies as such

an exclusion restriction, as we constructed all of them based on the assumption that they might

affect outcomes. We therefore followed recent research in financial economics (e.g., Acharya,


Davydenko, & Strebulaev, 2012) and generated several auxiliary interaction terms from the

available law firm variables that capture the notion of ‘herd behavior’: An acquirer is likely to

put more weight on an attribute as selection criterion, if more other acquirers rely on the same

criterion. We opted for law firm worldwide experience, interacted it with the variable’s average

among all lawyers appointed as legal counsel in the same country and same year as the focal

deal, and included the resulting interaction term in the selection equations.

RESULTS

Summary Statistics

Table 1 presents means, SDs, and Pearson correlation coefficients of all variables in

Sample B (see online appendix 2 for the corresponding table for the larger Sample C.) As the

variable range of many of the count measures is substantial, while their means and medians are

comparatively small, we computed the natural logarithms to ensure that our regression results are

not driven by a few firms with very large counts. Correlations between the mmm variables vary

between -.06 and .78, supporting the notion that they indeed describe knowledge structures in

different domains (Walsh, 1995). Also, the correlations between (a) the mmm variables and (b)

acquirer cross-border M&A experience and acquirer size, respectively, are—with a maximum of

P.C.<.81—in the range of values typical of experiential learning studies (e.g., Haunschild &

Sullivan, 2002).

--------------------------------------

Insert Table 1 about here

--------------------------------------

Acquirers’ Criteria for Selecting Legal Advisors


Table 2 presents the results of testing H1a and H2a. Model 1 (2) investigates the selection

criteria that acquirers apply in their choice of a legal counsel for a domestic (cross-border) deal.

--------------------------------------


--------------------------------------

H1a suggested that an advisor would more likely be selected for a domestic deal, if it had

more country expertise and closer connections with the acquirer. Indeed, advisor-client ties,

advisor-client tie strength, and acquirer country expertise are all significant and positive in

Model 1. We also posited that network-related criteria would have a stronger impact on

selection. As both advisor-client tie strength and acquirer country expertise are measured in

percentage points, we tested this formally by comparing their unweighted effect sizes. In

particular, as suggested by Wiersema and Bowen (2009), we compared their marginal effects.

The results showed that advisor-client tie strength alone is a more important selection criterion

than acquirer country expertise over the entire range of observations, i.e., acquirer-law firm

dyads in Sample C (p<.021). Thus, H1a is supported. H2a suggested that acquirers would

transfer selection rules from domestic to cross-border M&As using basically the same relative

weight of criteria, with prior ties being predominant. In Model 2, the coefficients of all four law

firm attributes are significant and similar in size to the ones of Model 1. Also, a test on the

relative size of marginal effects of advisor-client tie strength vs. the country expertise variables

shows that even tie strength alone is more important for selection than the two expertise-related

variables together (p<.001). H2a is therefore supported.


Determinants of Pre-Completion Outcomes

First, we check in how far sample selection is an issue and in how far our exclusion

restriction is a valid instrument. Our exclusion restriction, the ‘herd variable’, is included in the

selection equations in Models 1 and 2 (Table 2). The coefficient is highly significant (p<.001),

suggesting that acquirers indeed seem to follow other acquirers in their selection criteria,

supporting our case of a valid instrument. The final lines of Models 7-10 (Table 3) present the

coefficients of the Inverse Mills ratio constructed from the selection equations and included in

the corresponding outcome equations. The ratio is only significant in the control variable

specification for domestic deals (Model 7, p=.002). It is insignificant in all other models

(p>.438), suggesting that sample selection is not a major issue for the analysis, in particular not

for the cross-border models. Because Heckman selection models are prone to produce

inefficiently high standard errors, Models 8-10 present, as recommended (Cameron & Trivedi,

2005: 552), coefficients and p-values of a standard OLS regression.

--------------------------------------


--------------------------------------

H1b predicted that a lawyer’s country-specific expertise and strength of links with the

acquirer would reduce closing time in domestic M&As. According to Model 9, only advisor-

client tie strength has the hypothesized negative effect (p<.001). Acquirer country expertise has

the predicted sign but is insignificant; advisor-client ties has no meaningful effect. We also

hypothesized that prior relations would contribute relatively more to completion than country

expertise. This is confirmed in a Wald test on the relative strength of advisor-client tie strength

against acquirer country expertise (p=.016). Thus, H1b is supported for the effect of advisor-


client tie strength. H2b predicted a significant role of expertise and prior ties also for the

completion of cross-border deals, but, in contrast to H1b, with country expertise as the pre-

dominant factor. The first part of H2b is partly supported; the second part is fully supported:

Only acquirer country expertise and target country expertise have the predicted negative effects

on duration, whereby the latter shows a stronger effect (Model 10). The two relational variables

are insignificant. A Wald test on the relative strength of target country expertise vs. advisor-

client tie strength shows that the difference in effect sizes is statistically significant (p=.008).

H2b is supported, especially for a counsel’s target country expertise in cross-border deals.

In terms of economic impact, raising target country expertise from the average value of

0.21 in Sample C to the maximum of 1.0 reduces the closing phase of an international deal by

21%. At an average duration of 80 days, this corresponds to a reduction of ~17 days. The effect

of advisor-client tie strength in domestic deals is equally sizable: Duration decreases by 26% or

21 days. Both effects are in the range of values found for established determinants of pre-

completion outcomes (Wong & O’Sullivan, 2001). They are larger than, for example, the

difference between a private and a tender offer for an attempted cross-border deal (15 days) and

almost match the difference between an all-cash and a stock financed deal (26 days).

Testing for Learning and the Impact of an Acquirer’s Multiplicity of Mental Models

H3 postulated that the discrepancy between selection criteria and drivers of law firm

performance in cross-border deals would be curvilinear in bidders’ cross-border M&A

experience, with a turning point beyond which acquirers would predominantly consider country

expertise in their selection. This curvilinear effect is tested in Model 3 (Table 2) by means of

four interaction terms between acquirer cross-border M&A experience, the square of experience,

and the two key variables of our outcome models: advisor-client tie strength and target country


expertise. If such a curvilinear effect were to exist, we would expect that the moderating effect of

bidder’s cross-border M&A experience on advisor-client tie strength as a selection criterion

(EL1) was significantly larger than the moderating effect of bidder’s experience on target

country expertise (EL2). Also, we would expect the moderating effect of the square of bidder

experience on advisor-client tie strength (EL3) to be lower than the moderating effect of the

square of experience on target country expertise (EL4). Both conjectures are supported (Table

2). They can, moreover, be corroborated in a comparison of the true moderating effects, as

recommended by Wiersema and Bowen (2009): In more than 99% of our acquirer-lawyer dyads,

the moderating effects (ME) of experience are statistically significant (p<.05 across both

comparisons) and satisfy as predicted in H3: ME(EL1)-ME(EL2)>0 and ME(EL3)-ME(EL4)<0.

Hence, with sufficient experience, acquirers may overcome the initial bias in their international

selection decisions. Another question is whether the typical acquirer will ever accumulate

enough experience to reach the turning point. The corresponding marginal effects plot shows that

a transaction history of at least 13 cross-border deals is necessary to reach this point (see online

appendix 4, Figure A1). This is the privilege of 12.5% of the most frequent acquirers: For the

majority of firms in our sample, experience alone is thus insufficient.

Models 4-6 (Table 2) test whether bidders with a greater number of mental models reach

the turning point earlier, requiring less experience (H4). The results largely support this

hypothesis. First, the positive difference between the linear parts of the experiential learning

interactions, EL1 and EL2, prevails in Models 4-6, suggesting that, at low levels of experience,

acquirers lean increasingly more on advisor-client tie strength for selection. At high levels of

experience, in turn, acquirers put more weight on target country expertise for selection,

especially if they also exhibit greater multiplicity of mental models. This is suggested by the


positive difference between the coefficients on the mental model interaction terms (MM2-

MM1>0 in all three models). To confirm this formally, we also computed the true second-order

moderating effects (2ME) of our mmm variables. As the code is not by default implemented in

the statistical software package, we predicted the moderating effects manually based on the

STATA source code printed in the appendix of Wiersema and Bouwens (2009: 690). Online

appendix 5 presents our code and Table A4 an overview of the results. Online appendix 8

presents additional robustness checks. To summarize, the moderating effects of all three mmm

variables satisfy, as predicted in H4: 2ME(MM1)-2ME(MM2) < 0 for at least 90% of the

acquirer-law firm dyads in Sample C. Thus, the turning point seems contingent on an acquirer’s

multiplicity of mental models, i.e. multiplicity shifts down the critical number of M&As required

before an acquirer puts more weight on a lawyer’s target country expertise. The marginal effects

plots of Figure 1 illustrate this, too: The typical bidder with low multiplicity never reaches the

turning point, regardless of the amount of experience it accumulates as observed in our data. In

contrast, almost all acquirers with a high level of multiplicity reach the turning point (see Table

A5 in online appendix 6 for details). Finally, we assessed H4 in a sub-sample of deals announced

in 2010. We computed correlations between the number of TMT nationalities of the acquirers

and attributes of their counsels (see also online appendix 7, Table A6). In support of H4, TMT

nationalities is positively correlated with, and only with, target country expertise (P.C.=.29,

p<.001).

DISCUSSION

In the context of acquirers’ selection of legal advisors in M&As, this study, first,

examined whether firms transfer specific routines and rules from one context (domestic) to

another (cross-border); further, we analyzed whether such a transfer might result in an (initially)


suboptimal match (in terms of legal counsels’ performance) between the transferred routines and

the novel context. On this basis, the study investigated whether accumulation of M&A

experience would allow acquirers to overcome any such mismatch by revising their selection

criteria in cross-border deals. Such rebalancing, we argued, would require a categorical reversal

of the relative importance attached to the two types of law firm attributes studied—country

expertise vis-à-vis network relations. We suggested that acquirers’ experiential learning would

need to involve a shift from lower-level learning, which occurs within a given set of rules (Fiol

& Lyles, 1985), to higher-level learning that would entail adjusting these rules themselves.

Finally, this study examined whether acquirers featuring a greater number of distinct mental

models would enjoy advantages in terms of a specific aspect of their organizational learning, i.e.,

their ability to recognize earlier the need to switch to higher-level learning.

The empirical results of this study indicate that selecting lawyers for domestic and

international mergers indeed calls for distinct selection rules: In domestic deals, bidders

predominantly select their legal counsels based on prior network relations, which also dominate

as determinants of lawyers’ performance towards shortening time to completion. In cross-border

deals, acquirers tend to rely as well (at least initially) on lawyers with whom they share prior

business relationships. Yet, these law firms are less likely to be experts in the specific target

country environment. In line with extant social network studies, we thus find evidence in this

novel empirical context of a dark side of embeddedness, i.e., acquirers (at least initially) pay

insufficient attention to better matching partners outside their network. Acquirers appear to learn,

though, when they accumulate M&A experience. After counter-productive lower-level learning

at first, they reach a turning point (after a large number of deals) at which they categorically

reverse the relative weights of network- and expertise criteria. Finally, acquirers with a greater


number of mental models needed less experience before reaching this turning point. Overall, we

believe that our results have important implications for theory and practice.

Organizational learning research. First, conceptually and empirically, this study builds

on and extends research into components of the organizational context that are conducive to

organizational learning (e.g., Argote, 2013; Argote & Miron-Spektor, 2011; Fiol & Lyles, 1985).

In particular, it adds to research on organizational learning by investigating multiplicity of

mental models as a complementary facilitator of a specific aspect of firms’ experiential learning:

the ability to recognize the need to switch from lower- to higher-level learning. The ability to

identify when to engage in which type of learning has been highlighted as an important element

of organizations’ learning already by Fiol and Lyles (1985). To the extent that it is closely

related to reflection on the underlying learning processes itself, our results also speak to studies

that seek to establish foundations of a third type of learning such as ‘meta-learning’ (Visser,

2007) drawing, in particular, on Argyris and Schön (1996) (for overviews, see, Visser, 2007;

Tosey et al., 2012). The shift from lower- to higher-level learning that is central to this study

adds an outcome-oriented element to this concept. By empirically documenting such a shift, and

identifying factors conducive to it, we also contribute empirically to this literature, which is

largely conceptual in nature or rests on anecdotal evidence. Further, with respect to H4, the

notion that a firm’s level of diversification might be related to “skills for higher-level learning”

(Fiol & Lyles, 1985: 811) has been voiced earlier, although the specific mechanisms that could

give rise to such an association have mostly not been spelled out explicitly (e.g., Fiol & Lyles,

1985; related, Miller, 1996). For example, Fiol and Lyles (1985: 811) concluded their conceptual

study with questions for future research, one of them being whether “diversified firms have

better skills for higher-level learning than do single business firms”. Conceptually, we propose a


possible mechanism—multiplicity of mental models—that could give rise to such an effect and

empirically test the resulting prediction for a specific aspect of learning, i.e. recognition of the

need to shift to higher-level learning.

Our findings may also hold value for research at the interface of organizational learning

and diversification (e.g., Dujisters et al., 2012; Pennings, Barkema, & Douma, 1994). While the

dominant view in the literature on diversification is skeptical regarding its implications for firm

value (e.g., Lang & Stulz 1994; Wan et al., 2011), recent studies cautiously point to potential, yet

contingent benefits of diversification—for some firms, in certain contexts, and in relation to

particular tasks (e.g., Mackey, Barney, & Dotson, 2017; Tate & Yang, 2015). Our results add a

specific aspect of learning to the list of corporate activities that could be positively affected by

diversification, including as well a suggestion for a possible mechanism underlying this effect.

Social network research. This study adds to social network studies by focusing on how

experiential learning may provide a safeguard against the ‘dark side’ of strong network ties that

may lead organizations to pay insufficient attention to better matching partners outside their

network (e.g., Lee, 2013; Mitsuhashi & Min, 2016). Our results support this view by

documenting the contingent value of network relationships in a novel empirical context (e.g.,

Gulati, 1995; Yang et al., 2011). In addition, we show that firms appear to face significant

difficulties in overcoming the risks of over-embeddedness, as they need to accumulate

substantial experience before starting to pay due attention to local experts outside their networks.

Factors such as causal ambiguity and infrequency of experiences complicate organizational

learning in complex corporate domains (Argote, 2013). In the present context, a complementary

influence may impede bidders’ learning: strategic behavior of familiar partners. Past ties may

become a source of opportunistic exploitation, as the creation and presence of trust generates


precisely those conditions under which fraudulent behavior yields the highest payoffs

(Granovetter, 1985). As familiar counsels gain from exclusive relationships, they might,

strategically or unintentionally, obstruct acquirers’ learning by striving to exclude more

competent (local) experts from the counseling team.

M&A research. This study adds to literature on M&A completion (e.g., Chakrabarti &

Mitchell, 2016; Dikova et al., 2010; Muehlfeld et al., 2012; Zhou et al., 2016) by pointing to

legal advisors as a complementary influence. The explanatory power of the results, compared

with prior studies, implies that this type of factor—external advisors whose relevance likely

varies depending on the specific merger phase—is important to include in future studies.

Implications for practice. First, the results shed light on the value of experience for

learning how to overcome the dark side of embeddedness. They suggest that learning might be

impeded by external parties who strive to exclude competent (local) experts from their alliances

for strategic reasons—and acquirers should be aware of this. Second, positive effects for

organizational learning may result from the presence of a whole range of mental models

available in an organization. These findings, thus, point into a similar direction as other studies

such as Tate and Yang (2015) who argue that firms may reap productivity benefits from

redeploying labor internally to alternative uses—a type of human capital investment that likely

contributes to the diffusion of knowledge about multiplicity of mental models throughout the

organization. In a similar vein, Powell and Rhee (2016) emphasized the value of variance

(though in experience), among others due to greater caution with which decision-makers

interpret past experience. Embracing a healthy dose of skepticism as to the generalizability of an

organization’s own experience and keeping an open mind for alternative worldviews may enable

firms to more easily engage in rule-changing higher-order learning, if needed.


LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

As any study, this study has its limitations, which could be addressed in future research. A

first limitation relates to internal validity. Future research might complement our dependent

variable with other measures, such as, for example, completion time including search time

(related, Uzzi & Lancaster, 2004). Yet, the relevance of different types of external advisors (e.g.,

lawyers, investment banks; see, e.g., Hayward, 2003) varies across the M&A process. Outcome

variables and scope of influence of the type of advisor should be aligned. For example, while

legal advisors impact the speed of closing, they are not usually heavily involved in bargaining

about premiums. Still, future research should probe whether our conclusions extend to other

combinations of advisor-outcome variables. A second limitation pertains to our measures of

organizational learning and multiplicity of mental models, neither of which is directly observed.

It is standard procedure in much of the learning literature to use counts of cumulative experience

as input measures (for an overview, Argote, 2013). Still, future research should probe the

robustness of our results using alternative measures and settings, as our data do not allow us to

unambiguously identify the factors that drive the initial discrepancy between selection criteria

and antecedents of lawyers’ performance in international deals, and the sluggishness of

acquirers’ adaptation. Consequently, we cannot fully rule out alternative explanations: For

example, even if a familiar advisor cannot provide the same high service quality as a local

expert, the long-term advisor might charge lower fees (Uzzi & Lancaster, 2004) or be more

trustworthy in terms of the variability of outcomes (cf. Baum et al., 2005).

Further, operationalizing multiplicity of mental models and disentangling the time and

cross-sectional dimensions, i.e. addressing potential overlaps in the effects of experience and

mental models, remains a challenge, not least because experience constitutes a key ingredient of


the emergence and development of knowledge structures (Walsh, 1995). By using different

indicators across several organizational domains and levels, we sought to alleviate such

concerns. Yet, we acknowledge the difficulty of appropriately operationalizing multiplicity of

mental models, in particular given the aggregate and indirect nature of our data. While we have

done our best to control for experience, our proxies for multiplicity of mental models may still be

intertwined with experience. Therefore, our empirical results should be interpreted as indicative

of our theory at best. Future research could consider alternative proxies and, using smaller

samples, employ other research designs, such as in-depth qualitative studies (e.g., Kaplan, 2008)

and methods such as causal mapping (e.g., Markoczy, 2001) that have been developed for the

elicitation and quantification of knowledge structures. Such an approach would allow for

incorporating the micro-foundations of organizational cognition, i.e. the social dynamics that

influence the development of shared cognition within organizations (e.g., Kaplan, 2008;

Markoczy, 2001)—a dimension that is absent from this study. Follow-up research could then add

valuable insights by offering a multi-level perspective (cf. Eggers & Kaplan, 2013). For

example, inter-organizational differences at the micro-level in the degree and way that managers

come to hold congruent mental models might translate into distinct levels of the proposed

generalized awareness effect, despite the same number of mental models as assessed here, thus

accounting for some of the remaining variation in our data. Finally, future research could

consider, focusing on performance feedback (Greve, 2003), how a lawyer’s performance,

relative to the acquirer’s aspiration level, interacts with or is influenced by multiplicity of mental

models. Our data did not allow us to pursue such questions, but they emanate from this study and

we hope that they will foster scholarly conversation on organizational learning in complex

corporate domains.


REFERENCES

Acharya, V., Davydenko, S. A., & Strebulaev, I. A. 2012. Cash holdings and credit risk. Review

of Financial Studies, 22: 3572-3609.

Adair, W. L., & Brett, J. M. 2005. The negotiation dance: Time, culture, and behavioral

sequences in negotiation. Organization Science, 16: 33-51.

Argote, L. 2012. Organizational learning: Creating, retaining and transferring knowledge. New

York, NY: Springer Science & Business Media.

Argote, L., & Miron-Spektor, E. 2011. Organizational learning: From experience to knowledge.

Organization Science, 22: 1123-1137.

Argyris, C., & Schön, D. A. 1978. Organizational learning. Reading, MA: Addison-Wesley.

Argyris, C., & Schön, D. A. 1996. Organizational learning II. Reading, MA: Addison-Wesley.

Baker, W. E. 1990. Market networks and corporate behavior. American Journal of Sociology, 96:

589-625.

Barkema, H. G., & Schijven, M. 2008. Toward unlocking the full potential of acquisitions: The

role of organizational restructuring. Academy of Management Journal, 51: 696-722.

Baum, J. A., Rowley, T. J., Shipilov, A. V., & Chuang, Y. T. 2005. Dancing with strangers:

Aspiration performance and the search for underwriting syndicate partners. Administrative

Science Quarterly, 50: 536-575.

Bell, S. T., Villado, A. J., Lukasik, M. A., Belau, L., & Briggs, A. L. 2011. Getting specific

about demographic diversity variable and team performance relationships: A meta-analysis.

Journal of Management, 37: 709-743.

Boone, A. L., & Mulherin, J. H. 2007. How are firms sold? Journal of Finance, 62: 847-876.


Boone, C., Wezel, F. C., Witteloostuijn, A. v. 2013. Joining the pack or going solo? A dynamic

theory of new firm positioning. Journal of Business Venturing, 28: 511-527.

Bromiley, P., & Rau, D. 2016. Social, behavioral, and cognitive influences on Upper Echelons

during strategy process: A literature review. Journal of Management, 42: 174-202.

Cameron, A. C., & Trivedi, P. K. 2005. Microeconometrics: Methods and applications.

Cambridge, U.K.: Cambridge University Press.

Carpenter, M. A., & Frederickson, J. W. 2001. Top management teams, global strategic posture,

and the moderating role of uncertainty. Academy of Management Journal, 44: 533-545.

Chakrabarti, A., & Mitchell, W. 2016. The role of geographic distance in completing related

acquisitions: Evidence from U.S. chemical manufacturers. Strategic Management Journal, 37:

673-694.

Chung, S. A., Singh, H., & Lee, K. 2000. Complementarity, status similarity and social capital as

drivers of alliance formation. Strategic Management Journal, 21: 1-22.

Coates, J. C. 2001. Explaining variation in takeover defenses: Blame the lawyers. California Law

Review, 89: 1301-1422.

Cohen, W. M., & Levinthal, D. A. 1990. Absorptive capacity: A new perspective on learning and

innovation. Administrative Science Quarterly, 35: 128-152.

Coleman, J. S. 1990. Foundations of social theory. Cambridge, MA: Harvard University Press.

Corwin, S. A., & Schultz, P. 2005. The role of IPO underwriting syndicates: Pricing, information

production, and underwriter competition. Journal of Finance, 60: 443-486.

Cyert, R. M., & March, J. G. 1963. A behavioral theory of the firm. Englewood Cliffs, NJ:

Prentice-Hall.


Dahlin, K. B., Weingart, L. R., & Hinds, P. J. 2005. Team diversity and information use.

Academy of Management Journal, 48: 1107-1123.

DePamphilis, D. M. 2008. Mergers, acquisitions, and other restructuring activities (4th ed.). Los

Angeles, CA: Elsevier Academic Press.

Dikova, D., Rao Sahib, P., & Witteloostuijn, A. v. 2010. Cross-border acquisition abandonment

and completion: The effect of institutional differences and organizational learning in the

international business service industry. Journal of International Business Studies, 41: 223-

245.

Duijsters, G., Heimeriks, K. H., Lokshin, B., Meijer, E., & Sabidussi, A. 2012. Do firms learn to

manage alliance portfolio diversity? The diversity performance relationship and the

moderating effects of experience and capability. European Management Review, 9: 139-152.

Eggers, J. P., & Kaplan, S. 2013. Cognition and capabilities: A multi-level perspective. Academy

of Management Annals, 7: 295-340.

Faulconbridge, J., & Muzio, D. 2016. Global professional service firms and the challenge of

institutional complexity: ‘Field relocation’ as a response strategy. Journal of Management

Studies, 53: 89-124.

Fiol, C. M., & Lyles, M. A. 1985. Organizational learning. Academy of Management Review, 10:

803-813.

Flavell, J. 1987. Speculations about the nature and development of metacognition. In F. E.

Weinert & R. H. Kluwe (Eds.), Metacognition, motivation, and understanding: 21-29.

Hillside, NJ: Erlbaum.

Gargiulo, M., & Benassi, M. 2000. Trapped in your own net? Network cohesion, structural holes,

and the adaptation of social capital. Organization Science, 11: 183-196.


Gilson, R. J. 1984. Value creation by business lawyers: Legal skills and asset pricing. Yale Law

Journal, 94: 239-314.

Gilson, R. J., & Black, B. 1996. The law and finance of corporate acquisitions. New York, NY:

Foundation Press.

Ginsberg, A. 1989. Construing the business portfolio: A cognitive model of diversification.

Journal of Management Studies, 26: 417-438.

Glendening, M., Khurana, I. K., & Wang, W. 2016. The market for corporate control and

dividend policies: Cross-country evidence from M&A laws. Journal of International Business

Studies, 47:1106-1134.

Granovetter, M. 1985. Economic action and social structure: The problem of embeddedness.

American Journal of Sociology, 91: 481-510.

Greve, H. R. 2003. Organizational learning from performance feedback: A behavioral

perspective on innovation and change. Cambridge, U.K.: Cambridge University Press.

Gulati, R. 1995. Does familiarity breed trust? The implications of repeated ties on contractual

choice in alliances. Academy of Management Journal, 38: 85-112.

Gulati, R., Gargiulo, M. 1999. Where do interorganizational networks come from? American

Journal of Sociology, 104: 1439-1493.

Haleblian, J., & Finkelstein, S. 1999. The influence of organizational acquisition experience on

acquisition performance: A behavioral learning perspective. Administrative Science

Quarterly, 44: 29-56.

Hambrick, D. C., Davidson, S. C., Snell, S. A., & Snow, C. C. 1998. When groups consist of

multiple nationalities. Organization Studies, 19: 181-205.


Haspeslagh, P. C., & Jemison, D. B. 1991. Managing acquisitions: Creating value through

corporate renewal. New York, NY: Free Press.

Haunschild, P. R., & Sullivan, B. N. 2002. Learning from complexity: Effects of prior accidents

and incidents on airlines’ learning. Administrative Science Quarterly, 47: 609-643.

Hayward, M. L. 2003. Professional influence: The effects of investment banks on clients'

acquisition financing and performance. Strategic Management Journal, 24: 783-801.

Heckman, J. J. 1979. Sample selection bias as a specification error. Econometrica, 47: 153-161.

Heinz, J. P., Nelson, R. L., & Laumann, E. O. 2001. The scale of justice: Observations on the

transformation of urban law practice. Annual Review of Sociology, 27: 337-362.

Hitt, M. A., Dacin, T. M., Levitas, E., Arregle, J., & Borza, A. 2000. Partner selection in

emerging and developed market contexts: Resource-based and organizational learning

perspectives. Academy of Management Journal, 43: 449-467.

Howard, M. D., Withers, M. C., Carnes, C. M., & Hillman, A. J. 2016. Friends or strangers? It

all depends on context: A replication and extension of Beckman, Haunschild, and Phillips

(2004). Strategic Management Journal, 37: 2222-2234.

Hunter, W. C., & Jagtiani, J. 2003. An analysis of advisor choice, fees, and effort in mergers and

acquisitions. Review of Financial Economics, 12: 65-81.

Hylton, K. N., & Deng, F. 2007. Antitrust around the world: An empirical analysis of the scope

of competition laws and their effects. Antitrust Law Journal, 74.2: 271-341.

Jemison, DB, Sitkin, SB. 1986. Corporate acquisitions: A process perspective. Academy of

Management Review, 11(1): 145–163.

Jensen, M. 2003. The role of network resources in market entry: Commercial banks’ entry into

investment banking. Administrative Science Quarterly, 48: 466-497.


Kaplan, S. 2008. Framing contests: Strategy making under uncertainty. Organization Science,

19: 729-752.

Karsten, C. J., Malmendier, U., & Sautner, Z. 2014. M&A negotiations and lawyer expertise.

Working Paper. Available at SSRN 2576866.

Krishnan, C. N. V., Masulis, R. W. 2013. Law firm expertise and merger and acquisition

outcomes. Journal of Law and Economics, 56: 189-226.

La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 2008. The economic consequences of legal

origins. Journal of Economic Literature, 46: 285-332.

Lang, L. H. P, & Stulz, R. M. 1994. Tobin’s Q, corporate diversification, and firm performance.

Journal of Political Economy, 102: 1248-1280.

Lazega, E. 2001. The collegial phenomenon: The social mechanisms of cooperation among peers

in a corporate law partnership. Oxford, U.K.: Oxford University Press.

Lee, J. J. 2013. Dancing with the enemy? Relational hazards and the contingent value of repeat

exchanges in M&A markets. Organization Science, 24: 1237-1256.

Lim, B. C., & Klein, K. J. 2006. Team mental models and team performance: A field study of the

effects of team mental model similarity and accuracy. Journal of Organizational Behavior,

27: 403-418.

Mackey, T. B., Barney, J. B., & Dotson, J. P. 2017. Corporate diversification and the value of

individual firms: A Bayesian approach. Strategic Management Journal, 38: 322-341.

Marano, V., Arregle, J. L., Hitt, M. A., Spadafora, E., & van Essen, M. van. 2016. Home country

institutions and the internationalization-performance relationship: A meta-analytic review.



March, J. 1991. Exploration and exploitation in organizational learning. Organization Science, 2:

71-87.

Markoczy, L. 1997. Measuring beliefs: Accept no substitutes. Academy of Management Journal,

40: 1228-1242.

Markoczy, L. 2001. Consensus formation during strategic change. Strategic Management

Journal, 22: 1013-1031.

Miller, D. 1996. A preliminary typology of organizational learning: Synthesizing the literature.


Mitsuhashi, H., & Min, J. 2016. Embedded networks and suboptimal resource matching in

alliance formations. British Journal of Management, 27: 287-303.

Mizruchi, M. S., & Stearns, L. B. 2001. Getting deals done: The use of social networks in bank

decision-making. American Sociology Review, 66: 647-671.

Muehlfeld, K., Rao Sahib, P., & Witteloostuijn, A. v. 2012. A contextual theory of

organizational learning from failures and successes – A study of acquisition completion in the

global newspaper industry, 1981-2008. Strategic Management Journal, 33: 938-964.

Nadolska, A., & Barkema, H. G. 2014. Good learners: How top management teams affect the

success and frequency of acquisitions. Strategic Management Journal, 35: 1483-1507.

Nielsen, B. B., & Nielsen, S. 2013. Top management team nationality diversity and firm

performance: A multilevel study. Strategic Management Journal, 34: 373-382.

Nystrom, P. C., & Starbuck, W. H. 1984. To avoid organizational crises, unlearn. Organizational

Dynamics, 12: 53-65.

Officer, M. S. 2003. Termination fees in mergers and acquisitions. Journal of Financial

Economics, 69: 431–467.


Pennings, J. M., Barkema, H., Douma, S. 1994. Organizational learning and diversification.

Academy of Management Journal, 37: 608-640.

Powell, K. S., & Rhee, M. 2016. Experience in different institutional environments and foreign

subsidiary ownership structure. Journal of Management, 42: 1434-1461.

Powell, W. W., & Smith-Doerr, L. 1994. Networks and economic life. In S. Neil & S. Richard

(Eds.), Handbook of economic sociology: 368-402. Princeton, NJ: Princeton University Press.

Prahalad, C. K., & Bettis, R. A. 1986. The dominant logic: A new linkage between diversity and

performance. Strategic Management Journal, 7: 485-501.

Rouse, W. B., & Morris, N. M. 1986. On looking into the black box: Prospects and limits in the

search for mental models. Psychological Bulletin, 100: 349-363.

Rossi, S., & Volpin, P. F. 2004. Cross-country determinants of mergers and acquisitions. Journal

of Financial Economics, 74: 277-304.

Samuelson, W., & Zeckhauser, R. 1988. Status quo bias in decision-making. Journal of Risk and

Uncertainty, 1: 7-59.

Schraw, G., & Dennison, R. S. 1994. Assessing metacognitive awareness. Contemporary

Educational Psychology, 19: 460-475.

Senge, P. 1990. The fifth discipline: The art and science of the learning organization. NewYork,

NY: Currency Doubleday.

Serdar Dinc, I., & Erel, I. 2013. Economic nationalism in mergers and acquisitions. Journal of

Finance, 68: 2471-2514.

Shah, R. H., Swaminathan, V. 2008. Factors influencing partner selection in strategic alliances:

The moderating role of alliance context. Strategic Management Journal, 29: 471-494.


Simons, D. J., & Chabris, C. F. 1999. Gorillas in our midst: Sustained inattentional blindness for

dynamic events. Perception, 28: 1059-1074.

Spender, J. C. 1987. Industry recipes: An enquiry in the nature and sources of managerial

judgement. Oxford, U.K.: Blackwell.

Stinchcombe, A. L. 1990. Information and organization. Los Angeles, CA: University of

California Press.

Szulanski, G. 1996. Exploring internal stickiness: Impediments to the transfer of best practice

within the firm. Strategic Management Journal, 17: 27-43.

Tosey, P., Visser, M., Saunders, M. N. 2012. The origins and conceptualizations of ‘triple loop’

learning: A critical review. Management Learning, 43: 291-307.

Tate, G., Yang, L. 2015. The bright side of corporate diversification: Evidence from internal

labor markets. Review of Financial Studies, 28: 2203-2249.

Uzzi, B., & Lancaster, R. 2004. Embeddedness and price formation in the corporate law market.

American Sociological Review, 69: 319-344.

Very, P., & Schweiger, D. 2001. The acquisition process as learning process: Evidence from a

study of critical problems and solutions in domestic and cross-border deals. Journal of World

Business, 36: 11-31.

Visser, M. 2007. Deutero-learning in organizations: A review and a reformulation. Academy of

Management Review, 32: 659-667.

Walsh, J. P. 1995. Managerial and organizational cognition. Organization Science, 6: 280-321.

Wan, W. P., Hoskisson, R. E., Short, J. C., & Yiu, D. W. 2011. Resource-based theory and

corporate diversification: Accomplishments and opportunities. Journal of Management, 37:

1335-1368.


Weitzel, U., & Berns, S. 2006. Cross-border takeovers, corruption, and related aspects of

governance. Journal of International Business Studies, 37: 786-806.

White & Case. 2009. Worldwide antitrust merger notification requirements. Washington, DC:

White & Case.

Wiersema, M. F., & Bowen, H. P. 2009. The use of limited dependent variable techniques in

strategy research: Issues and methods. Strategic Management Journal, 30: 679-692.

Wong, P., & O’Sullivan, N. 2001. The determinants and consequences of abandoned takeovers.

Journal of Economic Surveys, 15: 145-186.

Yang, H., Lin, Z., & Peng, M. 2011. Behind acquisitions of alliance partners: Exploratory

learning and network embeddedness. Academy of Management Journal, 54: 1069-1080.

Zhou, C., Xie, J., & Wang, Q. 2016. Failure to complete cross-border M&As: “To” vs. “from”

emerging markets. Journal of International Business Studies, 47: 1077-1105.


FOOTNOTES

1 First, the public announcement presupposes that fundamental issues such as strategic fit have

mostly been resolved (Haspeslagh & Jemison, 1991). Second, announcing a deal entails the

revelation of private information about a bidder’s strategy (Officer, 2003), and disruption of

organizational routines (Jemison & Sitkin, 1986), and puts reputations on the line (Luo, 2005).

Third, Karsten et al. (2014) emphasize agency concerns: Acquirers may prefer a fast closing as

sellers retain control until official consummation, allowing them to extract private benefits.

2 Various factors may trigger a switch between levels of learning: first, factors external to the

learning process such as crises (e.g., Fiol & Lyles, 1985); second, factors inherent to the learning

process such as experience (e.g., Haleblian & Finkelstein, 1999). Third, performance feedback

theory emphasizes their joint influence by focusing on how performance is interpreted relative to

own historical and to social aspiration levels (e.g., Greve, 2003), predicting the existence of a

performance threshold beyond which organizations engage in distant rather than local search.

3 Thus, we assume a substantial degree of collectivity of these group mental models, rather than

investigating the dynamic and interactive social processes that influence the development of

shared cognition within organizations (e.g., Kaplan, 2008, 2011; Markoczy, 1997, 2001).

4 To improve data quality and validity of results, we manually corrected for typos in lawyer

names and country affiliations (by checking the websites), accounted for mergers in the industry,

and removed undisclosed acquirers and acquirers that represent nations or investor groups.

5 As alternatives to focusing on the largest advisor, we (1) analyzed only single-advised deals and

(2) reproduced the analysis at the advisor team level, by calculating the law firm-related variables

based on averages and maxima of the counsels’ attributes, respectively. Results (available upon

request) remained qualitatively unchanged.


Table 1

Summary Statistics of Sample B Pearson pairwise correlations

Variables Mean Std. Dev Min Max 1. 2. 3. 4. 5. 6. 7.

1. Completion time (log(days)) 3.0 1.9 0 8.1

2. Advisor-client ties (#) 1.6 3.5 0 48.0 .01

3. Advisor-client tie str. [0,1] .19 .29 0 1 -.09* .39*

4. Law firm acq. ctr. exp. [0,1] .46 .39 0 1 -.06* .17* .23*

5. Law firm target ctr. exp. [0,1] .33 .37 0 1 -.07* .07* .06* .42*

6. Law firm worldwide exp. (log(#)) 5.3 1.4 0 7.0 .12* .13* .08* -.14* -25*

7. Number of advisors (#) 1.7 1.3 1 12.0 .24* -.03 -.13* -.13* -.19* .29*

8. Acq. internat. M&A exp. (log(#)) 1.2 1.1 0 4.3 .08* .24* -.20* -.19* -.10* .09* .05*

9. (Acq. internat. M&A exp. (log)) 2 2.8 3.5 0 18.9 .07* .25* -.18* -.14* -.09* .09* .04*

10. Acq. network size (log(#)) 1.6 .92 0 4.0 .08* .28* -.19* -.03* -.04* .15* .07*

11. Acq. home ctr. exp. [0,1] .77 .32 0 1 -.04* .02 .19* .19* .13* -.03 -.01

12. Acq. target ctr. exp. [0,1] .14 .24 0 1 .01 .13* .18* .45* .50* -.06* -.06*

13. Acq. mmm antitrust systems [1,10] 2.6 1.6 1 9 .04* .22* -.15* -.14* -.09* .07* .02

14. Acq. mmm political systems [1,10] 3.3 2.0 1 10 .07* .21* -.17* -.15* -.11* .12* .05*

15. Acq. mmm product markets (log(#)) 2.1 .93 0 4.8 .08* .22* -.15* -.04* -.02 .06* .03*

16. Acq. nationalities TMT [0,1] .35 .25 .07 1 -.04* -.01 -.13 -.10* .15 .03 .02

17. Acq. size (log($US)) 9.2 2.4 -3.2 14.9 .18* .12* -.22* -.12* -.04* .08* .01*

18. Public acquirer (0,1) .74 .44 0 1 .04* -.05* .14* -.01 .01 -.01 .03*

19. Deal value (log($US)) 5.2 1.7 2.3 12.2 .45* .05* -.12* -.09* -.07* .25* .43*

20. Number of bidders (#) 1.04 .21 1 4 .09* -.02 -.04* -.01 -.01 .05* .12*

21. Public target (0,1) .30 .47 0 1 .37* -.01 -.08* -.02 .02 .08* .22*

22. Stock payment [0,1] .12 .31 0 1 .17* .03* .08* .13* .06* -.02 .11*

23. Tender offer (0,1) .14 .35 0 1 .21* -.03 -.06* -.07* -.02 .05* .15*

24. Common law ctr. (0,1) .70 .46 0 1 -.01 08* .08* .29* .54* -.01 -.02

25. Same legal origin (0,1) .98 .10 0 1 -.01 .06* .05* .20* .23* -.06* -.05*

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

9. (Acq. internat. M&A exp.) 2 .94*

10. Acq. network size .72* .71*

11. Acq. home ctr. exp. -.36* -.34* -.29*

12. Acq. target ctr. exp. -.18* -.17* -.08* .36*

13. Acq. mmm antitrust systems .81* .82* .64* -.34* -.17*

14. Acq. mmm political systems .79* .79* .70* -.33* -.16* .78*

15. Acq. mmm product markets .55* .63* .56* -.25* -.08* .53* .55*

16. Acq. mmm nationalities TMT -.08 -.03 -.12* -.35* -.06 -.06 -.01 .01

17. Acq. Size .58* .58* .51* -.32* -.12* .52* .52* .46* -.02

18. Public acquirer -.29* -.29* -.29* .44* .11* -.27* -.29* -.30* .24 -.26*

19. Deal value .18* .16* .17* -.05 .02 .13* .15* .13* .04 .31* .04*

20. Number of bidders .01 .01 .01 -.02 -.01 -.01 .01 .01 -.01 -.01 -.01*

21. Public target .06* .05* .03* -.05* .07* .03 .03 .06* .21 .01 -.01*

22. Stock payment -.11* -.09* -.13* .14* .19* -.09* -.14* -.11* -.10 -.06* .20*

23. Tender offer .05* .04* -.01 -.09* -.03* .03 .01 .04* .30 -.01 .04*

24. Common law ctr. -.12* -.11* -.04* .10* .40* -.13* -.10* .01 .19 -.08* .01

25. Same legal origin -.10* -.09* -.03* .17* .42* -.09* -.08* -.03* -.08 .04* .05*

19. 20. 21. 22. 23. 24.

20. Number of bidders .12*

21. Public target .18* .21*

22. Stock payment .18* .03 .24*

23. Tender offer .08* .17* .59* -.01

24. Common law ctr. -.01 .02 .09* .08* .01

25. Same legal origin .01 .01 .06* .11* .02 .31*

No. announced mergers = 11,511

NOTES: (a) * Significance level: p<.001. (b) Pairwise correlations for acquirer nationalities TMT only available for a sample of

65 law firm-merger dyads in 2010.


Table 2

Results of the Selection Models Selection equations: Probability of selecting a law firm

Expected

sign

Domestic

mergers

(1)

Internat.

mergers

(2)

Internat.

mergers:

experiential

learning

(3)

Internat.

mergers:

antitrust

systems

exposure

(4)

Internat.

mergers:

politcial

systems

exposure

(5)

Internat.

mergers:

Product

markets

exposure

(6)

Mental model second-order interactions

Mental model variable X

Experiential learning int.3 (MM1) (H4: -) -.02 (.072) -.01 (.317) -.10 (.000)

Experiential learning int.4 (MM2) (H4: +) .01 (.002) .01 (.000) -.01 (.529)

Experiential learning interactions

Acq. cross-border M&A exp. (log) X

Advisor-client tie strength (EL1) (H3: +) .31 (.002) .21 (.059) .20 (.059) .10 (.388)

Law firm target ctr. exp. (EL2) (H3: -) -.05 (.165) .02 (.685) .03 (.372) -.04 (.326)

(Acq. cross-border M&A exp. (log))2 X

Advisor-client tie strength (EL3)

(H3: -)

-.07 (.121)

.04 (.565)

.14 (.821)

.26 (.000)

Law firm target country exp. (EL4) (H3: +) .03 (.025) -.03(.170) -.03 (.052) .01 (.452)

Exclusion restriction

Law firm worldwide experience (log) X

Avg. law firm exp.in same country-year

(+)

.01 (.000)

.01 (.000)

.01 (.000)

.01 (.000)

.01 (.000)

.01 (.000)

Law firm variables

Advisor-client ties (H1: +) .06 (.000) .09 (.000) .09 (.000) .09 (.000) .09 (.000) .10 (.000)

Advisor-client tie strength (H1: +) 2.08 (.000) 2.02 (.000) 1.90 (.000) 1.90 (.000) 1.90 (.000) 1.88 (.000)

Acquirer country expertise (H1: +) .77 (.000) .20 (.000) .20 (.000) .20 (.000) .21 (.000) .20 (.000)

Target country expertise (H1: +) --- .66 (.000) .66 (.000) .65 (.000) .64 (.000) .65 (.000)

Worldwide experience (log) (+) 32 (.000) .21 (.000) .20 (.000) .21 (.000) .20 (.000) .21 (.000)

Number of advisors (+) .20 (.000) .17 (.000) .17 (.000) .17 (.000) .17 (.000) .17 (.000)

Acquirer variables

Cross-border M&A exp. (log) (-) .09 (.000) .02 (.031) .02 (.200) .02 (.193) .02 (.173) .02 (.347)

(Cross-border M&A exp. (log))2 (-) -.03 (.000) -.02 (.000) .02 (.001) -.01 (.004) -.01 (.001) -.02 (.002)

Network size (log) (?) -.10 (.000) -.02 (.006) -.02 (.017) -.02 (.015) -.02 (.009) -.02 (.014)

Home country experience (-) -.17 (.000) -.04 (.023) -.04 (.024) -.04 (.022) -.04 (.034) -.05 (.012)

Target country experience (-) --- -.16 (.000) -.16 (.000) -.16 (.000) -.15 (.000) -.16 (.000)

Acq. mmm antitrust systems (-) --- -.01 (.398) -.01 (.455) -.01 (.149) --- ---

Acq. mmm political systems (-) --- --- --- --- -.01 (.866) ---

Acq. mmm product markets (log) (-) --- --- --- --- --- .02 (.226)

Acquirer size (log) (?) -.01 (.475) .01 (.357) .01 (.376) .01 (.325) -.01 (.307) -.00 (.689)

Public acquirer (?) .01 (.319) -.01 (.609) -.01 (.555) -.01 (.587) -.01 (.780) -.01 (.348)

Announced mergers (sample B) 5,594 5,917 5,917 5,917 5,680 5,917

Law firm-mergers dyads (sample C) 524,506 634,669 634,669 634,669 613,482 634,669

NOTES: (a) The table reports point estimates and p-values in parentheses. (b) All models include a constant, 7 merger control

variables introduced in the methodology section, and 12 year dummies (not reported). (c) Models 4-6 omit two potential

interaction terms between the multiplicity of mental model variables and the linear experience interactions, EL1 and EL2. The

omission has solely a statistical reason, since these “linear” interaction terms are highly correlated with MM1 and MM2

(P.C.>.95). Nevertheless, from a logical viewpoint, either set of interaction terms is sufficient to test H4, because both shift the

turning point of the experiential learning parabola. The question still is whether our findings are robust to that choice. In online

appendix 8 Table A8, we successfully reproduce the results presented above using the alternative set of mental model interaction

terms.


Table 3

Results of the Outcome Models

Outcome equations: Duration of closing phase (log)

Expected

sign

Heckman:

Domestic

mergers

(7)

OLS:

International

mergers

(8)

OLS:

Domestic

mergers

(9)

OLS:

International

mergers

(10)

Law firm variables

Advisor-client ties (H2: -) .01 (.065) .01 (.530)

Advisor-client tie strength (H2: -) -.36 (.000) .01 (.889)

Acquirer country expertise (H2: -) -.05 (.512) -.21 (.018)

Target country expertise (H2: -) --- -.33 (.000)

Worldwide experience (log) (-) .02 (.311) .04 (.038) -.02 (.323) .02 (.369)

Number of advisors (+) -.01 (.885) .03 (.138) .01 (.998) .02 (.271)

Acquirer variables

Cross-border M&A exp. (log) (-) -.05 (.371) -.02 (.734) -.07 (.240) -.03 (.603)

(Cross-border M&A exp. (log))2 (-) .01 (.699) -.02 (.399) .01 (.594) -.02 (.380)

Network size (log) (-) .02 (.579) -.01 (.975) .01 (.843) .02 (.698)

Home country experience (-) -.03 (.767) -.14 (.148) -.03 (.718) -.10 (.282)

Target country experience (-) --- .02 (.826) --- .05 (.660)

Acq. mmm antitrust systems (?) --- -.03 (.209) --- -.03 (.194)

Acquirer size (log) (+) .03 (.022) .07 (.000) .03 (.036) .07 (.000)

Public acquirer (+) .14 (.013) .13 (.040) .15 (.011) .13 (.037)

Merger variables

Deal value (log) (+) .39 (.000) .41 (.000) .39 (.000) .41 (.000)

Number of bidders (+) .05 (.708) .01 (.934) .06 (.614) .02 (.909)

Public target (+) 1.08 (.000) .81 (.000) 1.08 (.000) .80 (.000)

Stock payment (+) .39 (.000) .38 (.000) .39 (.000) .38 (.000)

Tender offer (+) .02 (.814) .20 (.035) .02 (.817) .20 (.033)

Common law target country (-) -.10 (.138) -.11 (.045) -.10 (.156) -.01 (.977)

Same legal origin (-) --- -.030 (.544) --- -.05 (.276)

Announced mergers (Sample B) 5,594 5,917 5,594 5,917

Law firm-mergers dyads (Sample C) 524,506 634,669 524,506 634,669

Inverse Mills Ratio .13 (.001) .04 (.383) .49 (.124) -.33 (.536)

NOTES: (a) The table reports point estimates and p-values in parentheses. (b) The p-values of the Inverse Mills

Ratio stem from a Wald test for independence of the outcome equations (Models 7-10) from the corresponding

selection equations (Models 1 and 2). H0: Coefficient of the Inverse Mills Ratio=0. The H0 is not rejected for

Models 8-10. Thus, as Heckman selection models produce inefficiently high standard errors, we present, as

recommended by Cameron & Trivedi (2005: 552), the coefficients and p-values of a standard OLS regression. (c)

All models include a constant and year dummies (not reported). (d) Models 8 and 10 show that an acquirer’s

multiplicity of mental models, as measured by the variable mmm antitrust systems, has in itself no direct effect on

duration. This also holds for the other two mental model proxies mmm political systems and mmm product markets

(unreported regressions, available upon request).


Figure 1

The Moderating Effect of M&A Experience and Multiplicity of Mental Models

NOTES: (a) The panels present: ME(advisor-client tie strength) - ME(law firm target country expertise) on the likelihood of

selection, at different levels of acquirer cross-border M&A experience and multiplicity of mental models (mmm). ME stem from

Models 4-6 (Table 2) and are illustrated by quadratic prediction plots. (b) The values for the mmm variables are chosen so as to

highlight the emergence and inwards shift of the turning point. The bottom panel, e.g., shows that no turning point is in reach for

acquirers in the bottom 50% of mmm product markets. A turning point emerges for acquirers in the 6th decile and is shifted to the

left in the 8th decile.

.03

.04

.05

.06

.07

Marg

ina

l E

ffect(

tie s

treng

th)-

ME

(ta

rget ctr

. exp)

0 1 2 3 4Bidder cross-border M&A experience (log)

MMM Antitrust 3rd decile MMM Antitrust 4th decile

MMM Antitrust 6th decile

.04

.045

.05

.055

.06

.065

Marg

ina

l E

ffect(

tie s

treng

th)-

ME

(ta

rget ctr

. exp)


MMM Stability 6th decile MMM Stability 7th decile

MMM Stability 8th decile

.03

.04

.05

.06

.07

.08

Marg

ina

l E

ffect(

tie s

treng

th)-

ME

(ta

rget ctr

. exp)


MMM Products 5th decile MMM Products 6th decile

MMM Products 8th decile

selecting legal advisors in m&as: organizational … · selecting legal advisors in m&as 2...

Documents